Kevin Leahy

Provably Correct Persistent Surveillance for Unmanned Aerial Vehicles Subject to Charging Constraints

In this work, we present a novel method for automating persistent surveillance missions involving multiple vehicles. Automata-based techniques were used to generate collision-free motion plans for a team of vehicles to satisfy a temporal logic specification. Vector fields were created for use with a differential flatness-based controller, allowing vehicle flight and deployment to be fully automated according to the motion plans. The use of charging platforms with the vehicles allows for truly persistent missions. Experiments were performed with two quadrotors over 50 runs to validate the theoretical results.

Distributed information gathering policies under temporal logic constraints

We present an algorithm for synthesizing distributed control policies for networks of mobile robots such that they gather the maximum amount of information about some a priori unknown feature of the environment, e.g. hydration levels of crops or a lost person adrift at sea. Natural motion and communication constraints such as “Avoid obstacles and periodically communicate with all other agents”, are formulated as temporal logic formulae, a richer set of constraints than has been previously considered for this application. Mission constraints are distributed automatically among sub-groups of the agents. Each sub-group independently executes a receding horizon planner that locally optimizes information gathering and is guaranteed to satisfy the assigned mission specification. This approach allows the agents to disperse beyond inter-agent communication ranges while ensuring global team constraints are met. We evaluate our novel paradigm via simulation.

Temporal logic motion planning using POMDPs with parity objectives: case study paper

We consider a case study of the problem of deploying an autonomous air vehicle in a partially observable, dynamic, indoor environment from a specification given as a linear temporal logic (LTL) formula over regions of interest. We model the motion and sensing capabilities of the vehicle as a partially observable Markov decision process (POMDP). We adapt recent results for solving POMDPs with parity objectives to generate a control policy. We also extend the existing framework with a policy minimization technique to obtain a better implementable policy, while preserving its correctness. The proposed techniques are illustrated in an experimental setup involving an autonomous quadrotor performing surveillance in a dynamic environment.

Persistent surveillance for unmanned aerial vehicles subject to charging and temporal logic constraints

In this work, we present a novel method for automating persistent surveillance missions involving multiple vehicles. Automata-based techniques are used to generate collision-free motion plans for a team of vehicles to satisfy a temporal logic specification. Vector fields are created for use with a differential flatness-based controller, allowing vehicle flight and deployment to be fully automated according to the motion plans. The use of charging platforms with the vehicles allows for truly persistent missions. Experiments were performed with two quadrotors for two different missions over 50 runs each to validate the theoretical results.

Control in belief space with Temporal Logic specifications

In this paper, we present a sampling-based algorithm to synthesize control policies with temporal and uncertainty constraints. We introduce a specification language called Gaussian Distribution Temporal Logic (GDTL), an extension of Boolean logic that allows us to incorporate temporal evolution and noise mitigation directly into the task specifications, e.g. “Go to region A and reduce the variance of your state estimate below 0.1 m2.” Our algorithm generates a transition system in the belief space and uses local feedback controllers to break the curse of history associated with belief space planning. Furthermore, conventional automata-based methods become tractable. Switching control policies are then computed using a product Markov Decision Process (MDP) between the transition system and the Rabin automaton encoding the task specification. We present algorithms to translate a GDTL formula to a Rabin automaton and to efficiently construct the product MDP by leveraging recent results from incremental computing. Our approach is evaluated in hardware experiments using a camera network and ground robot.

Always choose second best: Tracking a moving target on a graph with a noisy binary sensor

In this work, we consider the problem of using a noisy binary sensor to optimally track a target that moves as a Markov Chain in a finite discrete environment. Our approach focuses on one-step optimality because of the apparent infeasibility of computing an optimal policy via dynamic programming. We prove that, under mild assumptions, always searching in the second most likely location minimizes one-step variance while maximizing the belief about the targets location over one step. Simulation results demonstrate the performance of our strategy and suggest the policy performs well over arbitrary horizons.

Integration of deterministic inference with formal synthesis for control under uncertainty

In this work, we consider an agent playing a turn-based game in a known environment against an adversary with unknown dynamics. The model of the adversary is assumed to belong to a subclass of regular languages that can be learned in the limit. We use tools from formal methods to synthesize a control strategy for the agent to win the game as it learns the model of its adversary, if a winning strategy exists. The strategy is updated as new information about the adversary is learned. The proposed framework is tested in simulation.

Semi-supervised pattern synthesis in spatially distributed dynamical systems

Pattern producing networks of dynamical systems are useful in a variety of applications ranging from formation control in multi-agent robotics to biological networks. We propose a formal methods framework with minimal user input to synthesize system parameters that result in the emergence of global steady state patterns in spatially distributed dynamical systems. Our framework consists of extensive exploration of steady state behaviors, dividing these behaviors into groups of patterns using a clustering algorithm, and determining system parameters for each pattern cluster using a recently developed formal methods approach for parameter synthesis. A case study illustrating the implementation of this framework on a network of locally interacting living cells is included.

Informative path planning under temporal logic constraints with performance guarantees

In this work we consider an agent trying to maximize a submodular reward function while moving in a graph environment. Such reward functions can be used to capture a variety of crucial sensing objectives in robotics including, but not limited to, mutual information and entropy. Furthermore, the agent must satisfy a mission specified by temporal logic constraints, which can encode many rich and complex missions such as “visit regions A or B, then visit C, infinitely often. Never visit D before visiting C.” We present an algorithm to maximize a submodular reward function under these constraints and provide an approximation for the performance of the proposed algorithm. The results are validated via simulation.

Localization of a Ground Robot by Aerial Robots for GPS-Deprived Control with Temporal Logic Constraints

In this work, we present a novel vision-based solution for operating a vehicle under Gaussian Distribution Temporal Logic (GDTL) constraints without global positioning infrastructure. We first present the mapping component that builds a high-resolution map of the environment by flying a team of two aerial vehicles in formation with sensor information provided by their onboard cameras. The control policy for the ground robot is synthesized under temporal and uncertainty constraints given the semantically labeled map. Finally, the ground robot executes the control policy given pose estimates from a dedicated aerial robot that tracks and localizes the ground robot. The proposed method is validated using a two-wheeled ground robot and a quadrotor with a camera for ten successful experimental trials.